Anti-piracy laws have come a long way since the advent of online streaming. But the fight against piracy has turned even more desperate in recent years as content, from TV shows to sports matches, is finding illegal ways into millions of households. Look at a couple of common examples, seasons of the TV show Game of Thrones have been pirated billions of times and football matches are streamed illegally to millions every day around the world.
Source: Irdeto
The relative ease with which a person can put a stream online is a major issue for these anti-piracy companies. The days of torrents have come and gone; all the person needs is a streaming box and an internet connection. By the time these anti-piracy firms launch their counter-attack and try to take down streams, it is usually always too late. Shutting down all available streams is an impossible task even with huge manpower allotted to the task.
This is where Irdeto has stepped in. Instead of throwing away hours of manual labor, Irdeto is using machine learning to fight back against the pirates. They have developed a convolution neural network that works through the internet to try to find illegal streams. Currently, the AI looks for broadcaster logos to identify which streams are illegal. The neural network was trained on a sample of more than a million images.
However, pirates are savvy folks and they have begun masking or blanking out logos from streams. This makes it incredibly hard to find illegal copies of the content. Irdeto is working on improving it’s current algorithm by making the machine understand how to distinguish between the kits each team wears. It’s also trying to incorporate facial recognition, wherein the machine will tag the face to the athlete/actor and enable a quick shutdown of the illegal course.
The firm is also hoping to use the video’s watermark to locate where the source of the stream is coming from. In response, the pirates either reduce the quality of the video so the watermark cannot be detected or use various sources to bring the stream together, rendering it almost untraceable.
The battle is almost relentless and as technology continues to grow, so will the fight between the anti-privacy firms and the pirate distributors.
Our take on this
One wonders if the people pirating the content will also start using advanced machine learning methods to circumnavigate companies like Irdeto? The hope is that more firms like Irdeto take up the fight against illegal streaming. But it goes to show the extent to which pirates will go to distribute the content and the incredible demand that sustains it.
Senior Editor at Analytics Vidhya.Data visualization practitioner who loves reading and delving deeper into the data science and machine learning arts. Always looking for new ways to improve processes using ML and AI.
